My professional journey and key experiences building impactful software and conducting research.
Machine Learning Research Intern
Dec 2025 - Present•Atlanta, Georgia, United States- ↳architecting KV cache stitching framework enabling 57% higher throughput and 44% more requests within TTFT/TPOT SLOs
- ↳engineering low-rank transformation layers with SVD initialization, reducing training loss by 35% across Llama & Qwen models
- ↳building distributed training pipeline using DeepSpeed & Accelerate achieving 0.7+ Rouge-L scores on NarrativeQA dataset
- ↳benchmarking 10K+ model configurations on H100 GPUs, measuring TTFT latency, TPOT improvements, and goodput gains
Software Engineer Intern
Jan 2025 - Dec 2025•Atlanta, Georgia, United States- ↳open-sourced NeurIPS & IEEE published research funded $15M by Meta & NSF, advancing Multivariate Time Series Forecasting
- ↳built inference dashboard with React.js & Tailwind CSS for Large Pre-Trained Time-Series Models (LPTMs) achieving 60ms latency
- ↳built benchmarking framework evaluating LPTMs vs Chronos & TimesFM using 16M+ datapoints from PostgreSQL & CSV pipelines
- ↳designed model serving infrastructure with 30+ REST API endpoints for dataset uploads, fine-tuning workflows, and real-time inference
Machine Learning Research Intern
May 2025 - Aug 2025•Athens, Georgia, United States- ↳Co-authored white papers on deep learning architectures for biomedical signal processing with novel CNN-BiLSTM hybrid model
- ↳architected 1D CNN classifier for signal quality assessment achieving 0.9917 F1 score via Pruning-based Bayesian Optimization
- ↳integrated CNN classifier as preprocessing head reducing BiLSTM model’s RMSE by 65% and eliminating training instabilities
- ↳engineered ETL pipeline loading ECG/PPG signals into PostgreSQL, applying FFTs & low-pass/high-pass filtering for noise removal
Teaching Assistant - CS106A
Apr 2025 - May 2025•Stanford, California, United States- ↳conducted weekly coding sessions, taught core Python concepts including control flow, data structures (lists, dictionaries), and object-oriented programming principles
- ↳graded assignments, distributed instructional material, and reinforced key programming concepts through structured exercises
- ↳throughout the 2 month program, attendance improved by 16.7% under my teaching
Machine Learning Research Intern
Jan 2025 - Apr 2025•Atlanta, Georgia, United States- ↳developed matrix addition, scalar multiplication, and matrix-vector operations in C++ and ported them to Kokkos for GPU-accelerated CFD solvers while working under Dr. Jain
- ↳conducted literature reviews and researched ML-based acceleration techniques for CFD, focusing on operator learning, super-resolution, and future flow state prediction using neural networks
- ↳wrote and tested PyTorch scripts to prototype models for real-time inference and to predict coefficients in the 1D Burgers' equation
Software Engineer Intern
May 2024 - Aug 2024•Wilmington, Delaware, United States- ↳built asynchronous FastAPI-based MCP server serving 50K+ concurrent requests with 100% uptime and OAuth2 authentication
- ↳engineered 25+ REST API endpoints transforming complex Apache Airavata API workflows into intuitive conversational interfaces
- ↳integrated LangChain AI Agent powered by open-source Qwen3 LLM into Apache’s $5M NSF-funded research platform, Cybershuttle
- ↳enabled natural language queries across 100K+ datasets, models, and notebooks at GPU-accelerated research clusters
Software Engineer Intern
Jan 2022 - Apr 2022•Dhaka, Bangladesh- ↳collaborated with SOLshare's Engineering team to design a portable solar panel delivering 25 watts of charging for up to 3 hours
- ↳prototyped solar solutions using Fusion 360, supporting SOLshare's mission to expand clean energy access
- ↳engineered circuit boards with soldering and wiring techniques to ensure device reliability
- ↳tested and validated products using Arduino IDE and C++ to meet safety and performance standards
